TU Darmstadt / ULB / TUbiblio

SmartFace: Efficient Face Detection on Smartphones for Wireless On-demand Emergency Networks

Lampe, Patrick and Baumgärtner, Lars and Steinmetz, Ralf and Freisleben, Bernd (2017):
SmartFace: Efficient Face Detection on Smartphones for Wireless On-demand Emergency Networks.
Limassol, Cyprus, In: 2017 24th International Conference on Telecommunications (ICT) (ICT 2017), Limassol, Cyprus, [Conference or Workshop Item]

Abstract

To support the search for missing persons during a natural disaster, photos taken by smartphone users staying inside the disaster area can be shared over wireless on-demand emergency networks formed by mobile devices. Detecting faces of persons in images and transmitting only the extracted faces can reduce the amount of transmitted data. However, executing common face detection algorithms on mobile devices is challenging, since these algorithms were not designed to cope with the devices' limited resources. In this paper, we present a novel approach to perform face detection locally on mobile devices in an efficient manner. The approach relies on a two-stage combination of existing face detection algorithms, enhanced by region of interest selection, color space/depth reduction, resolution scaling, face size definition, image scaling, image cropping, and bounding box scaling. Experimental results indicate that the proposed approach improves both the overall face detection rate and the overall runtime compared to the individual face detection algorithms used alone, and also reduces the amount of data that needs to be stored on disk and sent over the network.

Item Type: Conference or Workshop Item
Erschienen: 2017
Creators: Lampe, Patrick and Baumgärtner, Lars and Steinmetz, Ralf and Freisleben, Bernd
Title: SmartFace: Efficient Face Detection on Smartphones for Wireless On-demand Emergency Networks
Language: English
Abstract:

To support the search for missing persons during a natural disaster, photos taken by smartphone users staying inside the disaster area can be shared over wireless on-demand emergency networks formed by mobile devices. Detecting faces of persons in images and transmitting only the extracted faces can reduce the amount of transmitted data. However, executing common face detection algorithms on mobile devices is challenging, since these algorithms were not designed to cope with the devices' limited resources. In this paper, we present a novel approach to perform face detection locally on mobile devices in an efficient manner. The approach relies on a two-stage combination of existing face detection algorithms, enhanced by region of interest selection, color space/depth reduction, resolution scaling, face size definition, image scaling, image cropping, and bounding box scaling. Experimental results indicate that the proposed approach improves both the overall face detection rate and the overall runtime compared to the individual face detection algorithms used alone, and also reduces the amount of data that needs to be stored on disk and sent over the network.

Place of Publication: Limassol, Cyprus
Uncontrolled Keywords: Face Detection; Emergency Communication; Mobile Device; Image Processing
Divisions: DFG-Collaborative Research Centres (incl. Transregio)
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet > A: Construction Methodology
DFG-Collaborative Research Centres (incl. Transregio) > Collaborative Research Centres > CRC 1053: MAKI – Multi-Mechanisms Adaptation for the Future Internet > A: Construction Methodology > Subproject A3: Migration
Event Title: 2017 24th International Conference on Telecommunications (ICT) (ICT 2017)
Event Location: Limassol, Cyprus
Date Deposited: 05 Jul 2017 11:44
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item